Monday, January 9, 2017

Managed futures - should we be disappointed?

Managed futures, as measured by the SocGen index, finished negative for the year. Many would have thought this was an odd 12-month return performance given the events of the year. Let's list some of the big moves:  the large equity decline in the first quarter, the equity gains in the fourth quarter, the bond gains and subsequent fall through the year, the dollar move higher, the BREXIT event, the US presidential election, and the comeback in oil to name a few. We could go on with some of the minor markets, but the overall conclusion is that there were trends and there were some large moves.

Unfortunately, one of the major flaws with any analysis of managed futures is that investors can always, after the fact, point to some large moves, and ask why they have not been exploited. This is the ultimate hindsight bias for analysis of this strategy. There were trends based on large price moves noticed in charts; therefore, trend-followers should have made money. 

Nevertheless, should investors have expected better performance in 2016? We think it is hard to draw the conclusion that returns should have been better for two reasons. One, the SocGen index masked the strong gains by some managers. Two, the key events were actually relatively short-term, so many managers were not able to exploit the moves. 

First, an index is not the same as the performance of a manager. We often use an index as a shorthand measure for a diversified portfolio of managed futures, yet actual performance for individual managers may differ substantially. Looking at the IASG database and sorting by managers with more than $50 mm in AUM and who are classified as trend followers finds 62 managers as a sample. Dropping the top and bottom 5% as outliers, the data show a range between -13.77 and 9.57 through November with a median of .25. The median is better than the index returns, but more significant is the wide range of performance. There were clear winners and losers in 2016 with some managers actually doing very well. Others has poor performance. While this dispersion can always be expected, it does show that all trend-followers are not alike.

Second, the key events that would have allowed for strong crisis alpha were actually limited in scope. The large decline in equities lasted only a matter of weeks before the reversal began. Longer-term trend followers generated gains only to see them reversed over the next two months. The BREXIT event was a surprise with a strong short-term move and some unexpected trends. If a manager was not prepositioned for this surprise event, he did not make large profits. The same could be said for the US election. Nevertheless, the dollar rally and bond sell-off were already in place at the end of September as measure by simple moving averages. 

The bond and dollar moves could have been exploited by trend followers, but the equity move would have been harder to exploit. Even with the bond and dollar moves, large intra-year reversals would have limited the upside gains. I review of some simple modes suggests that better returns may have been possible, on average, but there was enough noise to cause wide dispersion based on simple model adjustments.

Systematic models capture specific price phenomena. If the environment does not exist, it does not matter what an investor may see in simple charts, money will not be made. Consequently, choosing a portfolio of managers is critical. However, while some diversification is valuable, there can be diminishing returns from too much diversification. Building a diversified portfolio is good, but too may managers may hurt overall gains.

No comments: